Class | Description |
---|---|
EvaluatorRequestSerializer |
Serialize and deserialize EvaluatorRequest objects
Currently only supports number & memory
Does not take care of Resource Descriptor
|
EvaluatorToPartitionMapper<V extends org.apache.hadoop.mapred.InputSplit> |
Class that tracks the mapping between
evaluators & the data partition assigned
to those evaluators.
|
InMemoryInputFormatDataSet<K extends org.apache.hadoop.io.WritableComparable<K>,V extends org.apache.hadoop.io.Writable> | |
InputFormatDataSet<K extends org.apache.hadoop.io.WritableComparable<K>,V extends org.apache.hadoop.io.Writable> |
An implementation of
DataSet that reads records using a RecordReader
encoded inside an InputSplit. |
InputFormatExternalConstructor |
A Tang External Constructor to inject the required
InputFormat
|
InputFormatLoadingService<K,V> |
An implementation of
DataLoadingService
that uses the Hadoop InputFormat to find
partitions of data & request resources. |
InputSplitExternalConstructor |
A Tang external constructor to inject an InputSplit
by deserializing the serialized input split assigned
to this evaluator
|
InputSplitExternalConstructor.SerializedInputSplit | |
JobConfExternalConstructor | |
JobConfExternalConstructor.InputFormatClass | |
JobConfExternalConstructor.InputPath | |
WritableSerializer |
A serializer class that serializes
Writable s
into String using the below Codec that
encodes & decodes Writable s
By default this stores the class name in the serialized
form so that the specific type can be instantiated on
de-serialization. |
Copyright © 2015 The Apache Software Foundation. All rights reserved.